Find answers from the community

Updated 11 months ago

My code is simple just following the

At a glance

The community member posted their code, which they say is simple and follows tutorials, but they are getting an empty response. A comment suggests that the community member can try to separate the retrieve and language model (LLM) steps, and enable debug logging to investigate the issue further. The comment provides example code to try retrieving data from the vector store and using the LLM to generate a response. There is no explicitly marked answer in the comments.

UUtine

My code is simple just following the tutorials, but I get empty response. This is my code:

Attachment

1 comment

AAshwinMS

I think internally query engine does a retrieve + llm step , i think u can try to do them individually and maybe even enable some debug logging @Utine ,

Plain Text

rr = index.as_retriever()
rr.retrieve("Some question which has revelant data in vec store")

and see its output. Then similarly can try for llm:

Plain Text

Settings.llm.predict("some relevant data from above + the original question u want to ask")

Add a reply